Synthesis Image Generation for Receipt Content Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems cannot effectively recognize the content of data being processed from evidence images that only include a data name, making it difficult to understand the processed image content.
Innovation Solution
An information processing apparatus that acquires a processed image and generates a synthesis image by combining an evidence image with a reduction image, such as a thumbnail image, to enhance content recognition, and adjusts the number of pages to fit within a receipt format, either by reducing individual pages or splitting the content across multiple synthesis images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the evidence image includes only a data name of the processed image, then the evidence image remains simple and compact, but the content of the processed image cannot be recognized
Solution Approach 1:
The patent combines the evidence image (containing processing information and fees) with a reduction image (thumbnail showing content preview) into a single synthesized evidence image. This merging allows both information completeness and content recognizability while maintaining a unified, manageable structure.
Solution Approach 2:
The reduction image acts as an intermediary element that bridges the gap between the simple evidence image and full processed image content. It provides a visual preview that enables content recognition without requiring the full high-resolution image, thus resolving the contradiction between simplicity and information completeness.
2Loss of information
If the processed image is included in full in the evidence image, then content recognition is improved, but the evidence image size and printing cost increase
Solution Approach 1:
Instead of providing the full high-resolution processed image, the patent uses a reduction image (thumbnail) that provides sufficient visual information for content recognition at a much smaller data size. This local quality approach delivers the necessary information quality without the full resource cost of the original image.
Solution Approach 2:
The patent creates a reduced-copy version of the processed image (thumbnail) that preserves the essential visual content characteristics while occupying minimal space. This copying strategy enables content recognition in the evidence image without incorporating the full high-resolution data, thus reducing data volume and printing costs.
3Productivity
If multiple processed images are synthesized in one evidence image, then information density increases, but the evidence image becomes too large for receipt format
Solution Approach 1:
The patent segments the synthesized evidence image into multiple separate evidence images when the total content exceeds receipt format capacity. Each segmented evidence image contains a subset of the processed images, ensuring that no single evidence image exceeds the printable area while maintaining high information density within each segment.
Data Source
AI summary
An information processing apparatus includes an acquisition unit that acquires a processed image obtained through image processing, and a generation unit that generates a synthesis image obtained by synthesizing an evidence image, which is an image becoming an evidence of the image processing and including at least information on a fee required for the image processing, with a reduction image obtained by reducing the processed image.